The central insight of the work is deceptively simple: because milk is produced by the mammary gland, its composition should carry information about the functional state of the organ that makes it. To test this idea, the researchers framed mammary gland function as three interconnected biological processes. The first is milk synthesis and secretion, the cellular machinery that manufactures milk components and releases them into the ductal system. The second involves the epithelial barrier and tissue remodeling, which govern what passes between the bloodstream and the milk while continuously reshaping the milk-producing tissue. The third encompasses immune and repair activity, which protects the gland from stress and restores tissue after injury. By mapping molecular patterns in milk onto these processes, the team asked not merely what is present in milk, but what that composition reveals about the gland producing it.

“Human milk gives us a rare, non-invasive window into the biology of the lactating mammary gland,” said Liat Shenhav, PhD, senior author of the study and an assistant professor at NYU Grossman School of Medicine with appointments in the Institute for Systems Genetics, the Department of Microbiology, and the Department of Obstetrics and Gynecology. “By looking at molecular patterns across diverse populations, we can begin to identify shared features of mammary function and understand how maternal nutritional and metabolic state relates to milk composition and infant growth.”

Despite the mammary gland’s central role in early life, how its functional state shapes milk composition, and how that variation relates to infant growth, has remained poorly understood. Human milk contains thousands of nutrients, metabolites, proteins, immune factors, and other molecules, but research has often examined these components one at a time rather than as an integrated biological system. The new study took a different approach. Across 1,543 human milk samples drawn from three cohorts in Canada, Pakistan, and Burkina Faso, the researchers combined multiple molecular layers, including metabolites, proteins, micronutrients such as vitamins, macronutrients such as carbohydrates, and human milk oligosaccharides, together with infant growth data and computational approaches including machine learning, a form of artificial intelligence.

The analysis drew on two randomized nutritional trials in which maternal supplementation promoted recovery or stability among infants at higher risk for poor growth. Multi-omic analyses, which integrate information across several molecule types such as proteins and metabolites to provide a complete picture of a biological system, revealed a shared pattern of milk composition associated with both maternal supplementation and infant growth. In other words, the same molecular signatures that shifted when mothers received nutritional support also tracked how well their infants were growing, suggesting a measurable molecular bridge between what mothers consume and how their babies develop.

The researchers then examined which molecules most consistently contributed to these shared nutrition-growth patterns. Their analysis identified seven candidate biomarkers linked to maternal nutritional status, infant growth, and biological programs related to mammary gland function. Four of the markers, C2, C3, C4, and C5, were short-chain acylcarnitines, molecules that help cells process nutrients for energy. The remaining three were lysine, an amino acid; 6′-sialyllactose, a human milk oligosaccharide; and selenium, a mineral. Together, the seven biomarkers captured complementary aspects of mammary biology, including milk synthesis, tissue remodeling, and immune response activity, and extended beyond established measures of blood-milk barrier permeability.

Short-chain acylcarnitines showed the strongest links across mammary functional domains, with C5 emerging as a particularly prominent signal. Higher C5 levels were associated with reduced milk synthesis signatures and with biological patterns related to tissue remodeling and immune activity. Because C5 stood out as a central metabolic signal, the researchers investigated which nutritional factors might explain variation in acylcarnitine levels. Acylcarnitine metabolism depends on coenzyme A, or CoA, a central molecule in energy and lipid metabolism that is synthesized from pantothenic acid, better known as vitamin B5. Drawing on associations across cohorts and established metabolic biology, the researchers propose that low vitamin B5 availability may constrain CoA-dependent metabolism and contribute to altered acylcarnitine profiles, including the accumulation of C5.

“Our findings demonstrate the intimate link between maternal nutrition, metabolic health and mammary function, highlighting the importance of an integrative approach to supporting maternal and infant health,” said first author Dr. April Jauhal of the Shenhav lab at NYU Grossman School of Medicine. “One plausible route linking maternal nutrition to mammary metabolism in our data is via the pantothenic-CoA-C5 metabolic pathway, and we are currently investigating the functional role of this pathway in mammary biology and lactation.”

The framing of milk as a biological window also places lactation within a broader continuum of maternal physiology that begins during pregnancy. During gestation, the mammary gland undergoes extensive hormonal differentiation, tissue and immune remodeling, and metabolic programming, while its major functional phenotype, lactation, becomes apparent only after delivery. In this sense, the mammary gland can be viewed as a postnatal maternal-infant interface: an organ programmed during pregnancy that subsequently connects maternal biology and nutrition with the developing infant. Like the placenta during pregnancy, it sits at a critical interface between mother and child, but human milk offers something the placenta cannot, a way to study that interface repeatedly and non-invasively after birth. Because milk samples can be collected over time, researchers can track how mammary function evolves across lactation without invasive procedures.

Building on this work, the Shenhav Lab is developing an integrated computational and experimental framework to test potential nutritional bottlenecks that may constrain mammary metabolism and milk production, including the vitamin B5-coenzyme A-C5 pathway. The lab is also studying colostrum and transitional milk to understand how mammary functional states emerge during the shift from pregnancy to mature lactation, and whether early molecular changes in milk reflect the emergence of mammary metabolic, barrier, and secretory function. The study is one of two complementary papers published in Science arising from the International Milk Composition, or IMiC, Consortium, an international research effort conducted in collaboration with Meghan Azad of the University of Manitoba and funded by the Gates Foundation.

The IMiC Consortium brought together more than 1,000 mother-infant pairs and profiled more than 25,000 molecular features in human milk. While the NYU Grossman-led study asked what milk composition can reveal about mammary gland function and its relationship with infant growth, the companion study asked a complementary question: which components of human milk can be changed through maternal nutrition? That companion study found that supplementation during lactation increased several B vitamins in milk, while many other milk components remained comparatively stable. Together, the two papers aim to determine how maternal nutritional and metabolic state during pregnancy and postpartum shapes mammary function, how that state is reflected in milk, and how these processes relate to infant development.

“The studies show that human milk is neither fixed nor simply a reflection of what a mother eats,” said Dr. Shenhav. “Some components respond to maternal nutrition, while others remain stable or appear to reflect broader biological states within the mammary gland. Understanding that distinction is essential if we want to design interventions that meaningfully support maternal and infant health.” The distinction matters practically as well as scientifically: if certain milk components are responsive to diet while others are buffered by gland biology, nutritional interventions can be targeted at the molecules most likely to change, rather than applied indiscriminately.

“Ultimately, we want to understand milk in the context of the biology that produces it,” said Dr. Shenhav. “The mammary gland sits at a critical interface between maternal physiology and the developing infant. By understanding how maternal metabolism influences mammary function and milk composition, we can begin to uncover the biological pathways linking maternal nutritional status with infant growth.” The work was funded by the Gates Foundation, the National Institutes of Health, the Canadian Institutes of Health Research, and a range of additional public and private supporters. For a field that has often treated milk as a static list of ingredients, the study marks a conceptual shift toward viewing it as a dynamic, system-level readout of one of the most remarkable organs in human physiology.

Subject of Research: Molecular patterns in human milk linking maternal nutrition, mammary gland function, and infant growth

Article Title: Human milk offers a window into mammary gland function—and its links to infant growth

Article References: Human milk offers a window into mammary gland function—and its links to infant growth. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: human milk, mammary gland, lactation, maternal nutrition, infant growth, acylcarnitines, vitamin B5, coenzyme A, multi-omics, human milk oligosaccharides, machine learning, biomarkers

The significance of coenzyme A in cardiac biology is difficult to overstate. This small, vitamin B5-derived cofactor sits at the crossroads of virtually every energy-generating pathway in the cell, shuttling acyl groups into the mitochondria for fatty acid oxidation, feeding the tricarboxylic acid cycle, and enabling the synthesis and breakdown of countless metabolites. Without adequate coenzyme A, the heart’s metabolic machinery grinds toward inefficiency. In the failing heart, a condition long recognized as one of profound metabolic inflexibility, the supply and utilization of this cofactor become compromised, leaving cardiomyocytes unable to fully exploit the fuels available to them. The failing heart, in this sense, is not merely an overworked pump but an energy-starved engine running on a degraded fuel system.

The pivotal evidence comes from a study by Forelli and colleagues published in Science in 2026, which demonstrated that SGLT2 inhibitors activate pantothenate kinase in the human heart. Pantothenate kinase is the first and rate-limiting enzyme of the coenzyme A biosynthesis pathway, the molecular gatekeeper that controls how much coenzyme A the cell can manufacture from pantothenate, the dietary precursor better known as vitamin B5. By stimulating PANK1, the drugs effectively reopen this gate, boosting coenzyme A availability and, with it, the myocardial metabolic capacity that heart failure erodes. Riccio and Santulli emphasize that PANK1 should now be considered a direct cardiac target of SGLT2 inhibitors, a designation that reframes how the field thinks about these drugs entirely.

What makes this finding particularly provocative is its independence from the drug’s canonical target. SGLT2, the glucose transporter in the kidney that gives the class its name, is barely expressed in cardiomyocytes, which has long puzzled researchers trying to explain why the heart benefits so much. Recent work has reinforced this disconnect: a 2024 study in Circulation Research showed that SGLT2 inhibitors confer benefit in mouse models of heart failure with reduced ejection fraction even in the absence of SGLT2, and a 2026 study in Metabolism identified early metabolic cardiac targets of empagliflozin that operate independently of SGLT2 inhibition. The PANK1–coenzyme A axis now emerges as one of the most concrete candidates for these SGLT2-independent actions, offering a direct biochemical route from drug to heart cell.

The commentary also situates the new discovery within a broader and evolving understanding of how SGLT2 inhibitors reprogram cardiac metabolism. Beyond glucose lowering, the class has been shown to enhance ketogenesis through an unexpected mechanism: rather than simply raising circulating ketone levels by shifting renal handling of glucose, the drugs act as allosteric activators of HMGCS2, the mitochondrial enzyme that catalyzes the rate-limiting step of ketone body synthesis, as reported in the Journal of Clinical Investigation in 2026. Meanwhile, the randomized, controlled EMPA-VISION trial assessed cardiac energy metabolism in patients with heart failure taking empagliflozin and provided a rigorous human window into how the drug reshapes myocardial energetics. Together, these lines of evidence paint a picture of a drug class that acts as a broad-spectrum metabolic modulator rather than a single-target agent.

Viewed through this lens, the cardioprotective portfolio of SGLT2 inhibitors begins to look like a coordinated campaign on multiple metabolic fronts. The PANK1 activation restores the coenzyme A supply needed to import and oxidize fatty acids, the heart’s preferred fuel. The HMGCS2 activation boosts ketone production, providing an oxygen-efficient alternative substrate. Additional studies, including work in frail hypertensive and diabetic patients published in Hypertension, have documented improved endothelial function and reduced mitochondrial oxidative stress with empagliflozin, extending the metabolic story to the vasculature. Each mechanism addresses a different facet of the energetic deficit that defines the failing heart, and together they may explain why the clinical benefits of these drugs have been so consistent across patient populations with and without diabetes.

The therapeutic implications of targeting the PANK1–coenzyme A axis extend well beyond explaining an existing drug class. If restoring coenzyme A biosynthesis can rescue myocardial metabolic capacity, then PANK1 itself becomes a druggable node for heart failure therapy, potentially separable from SGLT2 inhibition altogether. This matters for patients who cannot tolerate SGLT2 inhibitors or who face contraindications, and it opens the door to rational combination strategies in which metabolic restoration is pursued deliberately rather than as a side effect of glucose handling. It also invites a re-examination of nutritional and pharmacological approaches to pantothenate metabolism in cardiac disease, a corner of cardiology that has received comparatively little attention despite the centrality of coenzyme A to heart function.

There are, of course, important questions that remain before the laboratory insight can be translated into bedside practice. The precise dose–response relationship between SGLT2 inhibitor exposure and PANK1 activation in the human heart, the time course over which coenzyme A restoration translates into functional improvement, and the extent to which this mechanism contributes to benefit relative to hemodynamic and neurohormonal effects all require further study. Human trials such as EMPA-VISION have begun to map the metabolic consequences of treatment in patients, but directly interrogating the PANK1 pathway in the myocardium remains a technical challenge. Riccio and Santulli’s commentary serves as both a synthesis of the current evidence and a roadmap for the experiments that will determine whether the coenzyme A story holds up as a central pillar of cardioprotection.

For a field that has spent two decades chasing the secret behind one of medicine’s most serendipitous successes, the identification of a direct cardiac target offers a rare moment of clarity. The heart fails, in large part, because it runs out of usable energy, and a drug class that replenishes the cofactor at the very heart of energy metabolism addresses the disease at its biochemical root. Whether PANK1 activation proves to be the dominant mechanism or one strand in a richer web of metabolic actions, the convergence of evidence from human hearts, mouse models, and clinical trials suggests that the era of thinking about SGLT2 inhibitors as mere glucose-lowering agents is definitively over. What began as a diabetes treatment has become a master key to cardiac metabolism, and the PANK1–coenzyme A axis may be one of its most important locks.

Subject of Research: Cardioprotective mechanisms of SGLT2 inhibitors through PANK1-mediated coenzyme A restoration in heart failure

Article Title: Beyond glucose lowering and ketogenesis: unveiling new cardioprotective mechanisms of SGLT2 inhibitors

Article References: Riccio, A., & Santulli, G. (2026). Beyond glucose lowering and ketogenesis: unveiling new cardioprotective mechanisms of SGLT2 inhibitors. Nature Reviews Cardiology. https://doi.org/10.1038/s41569-026-01357-w

Image Credits: AI Generated

DOI: 10.1038/s41569-026-01357-w

Keywords: SGLT2 inhibitors, heart failure, PANK1, coenzyme A, cardiac metabolism, pantothenate kinase, ketogenesis, empagliflozin, HMGCS2, mitochondria, cardioprotection, metabolic flexibility

The investigation began with an unbiased metabolomic survey of mice with chemically induced colitis. Rather than confirming the team’s expectations, the data pointed somewhere unexpected: levels of pantothenate, the vitamin B5-derived precursor of coenzyme A, were markedly elevated in both the colon and serum of colitic animals. Because pantothenate must be phosphorylated by pantothenate kinases, known as PANK enzymes, before it can enter the coenzyme A biosynthetic pathway, its accumulation signaled that the pathway itself was stalled. Of the three PANK isoforms present in the colon, one stood out. PANK3, the dominant intestinal form, was dramatically reduced at both the messenger RNA and protein levels in colitic mice, and the drop was confirmed by immunohistochemical staining of the colonic epithelium.

The downregulation was not merely a rodent curiosity. Analyzing gene expression datasets from human intestinal biopsies, the researchers found PANK3 significantly decreased in patients with active Crohn’s disease and active ulcerative colitis compared with inactive disease and healthy controls. Tissue staining of patient biopsies from Jiangsu Province Hospital corroborated the finding, showing diminished PANK3 specifically within the intestinal epithelium. When human colon epithelial cell lines were exposed to the inflammatory cytokines TNF-alpha and IFN-gamma, PANK3 levels fell again, replicating in a dish what had been observed in diseased tissue. Taken together, the evidence positioned epithelial PANK3 deficiency as a consistent pathological feature of inflammatory bowel disease across species.

To determine whether the loss of PANK3 actually drives barrier failure rather than simply accompanying it, the team deployed a combination of pharmacological and genetic tools. In cultured epithelial cells subjected to inflammatory challenge, a small-molecule PANK3 agonist called PZ-2891 boosted the expression of intercellular junctional proteins, while Hopantenate, a competitive PANK inhibitor, worsened their decline. In Caco-2 monolayer models, transepithelial electrical resistance and paracellular permeability assays showed that activating PANK3 restored barrier function while blocking it deepened the damage. Directly overexpressing PANK3 in HT29 and NCM460 cells rescued inflammation-induced barrier injury, whereas silencing the gene produced the opposite effect, confirming the relationship in both directions.

In living animals the results were even more striking. Mice engineered to overexpress PANK3 specifically in the intestine via adeno-associated virus were markedly protected against DSS-induced colitis, exhibiting less weight loss, shorter reduction in colon length, preserved crypt architecture, maintained goblet cell populations and reduced inflammatory infiltration. Serum FITC-dextran assays demonstrated tighter barrier integrity, and transmission electron microscopy revealed that shortened microvilli and disrupted apical junction complexes were structurally repaired in the PANK3-overexpressing animals. Conversely, intestinal PANK3 knockdown exacerbated every measure of disease severity and barrier breakdown, establishing PANK3 as a necessary guardian of the epithelial lining during colonic inflammation.

The mechanism connecting a metabolic enzyme to such profound structural changes emerged from transcriptomic profiling. PANK3 overexpression suppressed gene clusters encoding proinflammatory cytokines, extracellular matrix components and drivers of epithelial-to-mesenchymal transition, or EMT, a developmental program in which epithelial cells lose their junctions and polarity and take on migratory, mesenchymal characteristics. In chronic inflammation, pathological persistence of EMT dismantles tight junctions and adherens junctions, degrades the basement membrane and promotes fibrotic remodeling, a signature observed in clinical IBD mucosal samples and correlated with disease severity and therapeutic resistance. PANK3, the study found, acts as a powerful brake on this process, downregulating mesenchymal markers such as vimentin and N-cadherin, suppressing the EMT-driving transcription factors Snail and Twist, and preserving junctional proteins including ZO-1, E-cadherin, occludin and claudins.

Gene set enrichment analysis revealed that the EMT-suppressing effects of PANK3 operated through inhibition of the PI3K/AKT and MAPK signaling pathways, both well-established inducers of the transition. Phosphorylation of PI3K, AKT and ERK fell sharply in PANK3-overexpressing mice and rose in knockdown animals, indicating the enzyme’s influence extended deep into canonical signal transduction. The critical mediator proved to be coenzyme A itself. Quantitative LC-MS/MS measurements showed that PANK3 overexpression significantly increased colonic free CoA abundance while knockdown depleted it, and supplying exogenous CoA to cultured epithelial cells was sufficient to suppress inflammation-induced EMT and downstream signaling. When pantothenate was completely removed from the culture medium, the barrier protection conferred by PANK3 agonism or overexpression was largely abolished, confirming that the enzyme’s benefit depends on the raw material for CoA synthesis.

Having established what PANK3 does, the team turned to why it disappears during colitis. Screening transcription factor prediction databases against the Pank3 promoter uncovered c-Myc, the notorious proto-oncogene, as the prime candidate. c-Myc was significantly upregulated in the colonic lesions of ulcerative colitis patients and in multiple experimental colitis models, and chromatin immunoprecipitation confirmed that c-Myc binds directly to three sites in the Pank3 promoter region. A dual-luciferase reporter assay demonstrated that c-Myc overexpression dramatically represses Pank3 promoter activity, and pharmacological inhibition of c-Myc with the inhibitor 10058-F4 restored PANK3 expression in a dose-dependent manner in cells. In mice, c-Myc inhibition ameliorated DSS-induced colitis, rescued PANK3 levels and improved barrier integrity, establishing a clear c-Myc-PANK3-EMT axis in which inflammation-driven c-Myc overexpression silences PANK3, depletes coenzyme A, permits EMT and dismantles the epithelial barrier.

Perhaps the most clinically tantalizing finding came from a high-throughput virtual screen for PANK3 agonists, which identified folic acid, a widely available and inexpensive B vitamin, as the top candidate. Molecular docking predicted binding to key residues including Lys24, Ser192, Arg207, Val268, Asn299 and Trp341, and a cellular thermal shift assay confirmed that folic acid directly stabilizes the PANK3 protein. Oral folic acid at 30 milligrams per kilogram significantly ameliorated DSS-induced colitis in mice, restoring body weight, colon length, crypt architecture and goblet cell populations while elevating colonic coenzyme A levels. Electron microscopy showed repaired microvilli and restored apical junction complexes, junctional proteins were upregulated, EMT markers and matrix metalloproteinases declined, and PI3K/AKT and MAPK activation was suppressed. Crucially, the protection vanished when PANK3 was knocked down, demonstrating that folic acid’s benefit is PANK3-dependent.

The findings carry substantial implications beyond the immediate identification of a druggable target. A meta-analysis cited in the study indicates that higher folate levels are associated with reduced risk of inflammatory bowel disease, lending clinical plausibility to the mechanistic work, although the therapeutic doses used in mice exceed typical supplementation levels and would likely require colon-targeted formulations such as enteric-coated tablets to maximize local exposure while limiting systemic dose. The authors also suggest that circulating pantothenate and coenzyme A levels, together with intestinal PANK3 expression, could serve as accessible biomarkers for diagnosis and treatment monitoring, pending validation in larger multi-center cohorts. Because reduced PANK3 has previously been identified as a diagnostic marker for early-stage colorectal cancer, and c-Myc is a well-established oncogene, the axis also offers a plausible mechanistic thread connecting chronic colitis to malignant transformation. Open questions remain, including precisely how PANK3-derived coenzyme A suppresses the PI3K/AKT and MAPK pathways, whether coenzyme A acts through post-translational modifications such as protein CoAlation, and whether PANK3 possesses non-catalytic functions. Even so, the study reframes ulcerative colitis in part as a metabolic disease of the epithelial barrier and positions PANK3 agonism, potentially through a humble vitamin, as a promising avenue for restoring the gut’s broken wall.

Subject of Research: Metabolic regulation of intestinal epithelial barrier integrity by the c-Myc-PANK3-coenzyme A axis in ulcerative colitis.

Article Title: c-Myc-PANK3-EMT axis regulates the structure and function of intestinal barrier in ulcerative colitis

Article References: Zhang, S., Chen, Y., Aa, N., Xu, C., Xie, T., Wang, Y., Cheng, T., Wang, M., Yu, H., Ji, X., Zhao, S., Wang, Y., Xiao, J., Xie, Y., Wang, G., & Aa, J. (2026). c-Myc-PANK3-EMT axis regulates the structure and function of intestinal barrier in ulcerative colitis. Journal of Advanced Research, 87, 931-946. https://doi.org/10.1016/j.jare.2025.12.007

Image Credits: AI Generated

DOI: 10.1016/j.jare.2025.12.007

Keywords: ulcerative colitis, PANK3, c-Myc, coenzyme A, epithelial-mesenchymal transition, intestinal barrier, folic acid, pantothenate kinase, inflammatory bowel disease, PI3K/AKT pathway, MAPK pathway, tight junctions